The Economics of Sex Differentials, 1974

Primary tabs

Download: Transcript

Pages: 1 2 3 4 5 6 7 8 9 View All

Show transcript

THE ECONOMICS OF SEX DIFFERENTlALS* ' K by Cynthia B. Lloyd and Beth T. Niemi \ . 1 I

There has been a great deal of debate recently on the sources of the wage and unemployment differentials that exist between men and women. Our major concern here is to try to sort out the relative importance of discrimination in explaining these two differentials, with particular emphasis on the new feminist perspective which is developing within the field of economics.

One school of economic andgysis believes it can explain these differentials in terms of differences between the sexes in their natural preferences for different kinds of work and in their productivity on the job. These

differences are believed to stem from women's dual role as a producer of

~non-market goods and services in the home and a worker in the labor force,

as contrasted with men's assumed single-minded commitment to market work. It is true that, on the average, women have chosen to work in,

less continuously than men. Earnings do increase with work experience in

general because both formal training and infonnal learning on the job increase:.H!‘

worker productivity. with less labor market experience and more movement'inl‘ and out of the labor force, women tend to be employed in inferior jobs at l~ lower wages, and also to suffer from more unemployment than men. Even so,

this approach still leaves much of the existing wage and unemployment differentials unexplained. T f _____;c;;tfil;_s:—fl;;a_;;a_séth.1; Niemi are Assistant Professors of-Economics

at Barnard College and Newark College of Arts and Sciences, Rutgers University, respectively. This paper evolved from the discussion which took place in a workshop chaired by the authors at the Barnard College conference on “The Scholar

and the Feminist," May 11, 1974, The authors wish to express their

appreciation to Harriet Zellner for her very helpful comments and suggestions.




K '! , ‘I _.\ .

.4 N . .-.p.....».a.(.'..._,,~:.' aw ..\.a.s. ....-I ; -.L"r:A—.\<-p «n«-~m.r- ... ...vr_._......<.«...-_«,.—x....-.. ..(,.e,,-,.,.,.,\..,. ',»_‘.~41r‘~m.'\.".'lA,\.vq./‘,\-[.)‘...~- 2- ;.,., ._-».'.i4.,..»a-a»,-u;....,;,:.


The qu«~.~:Linn l‘e‘.!11.'"1l11!~§ as to whether the unexplained portion of these

i _ , I - ‘ » \ differentials is the result of discrimination. Those who have attempted to explain these sex differentials by means of the above analysis suggest that better measurement of the relevant variables will enable them to explain all of the wage and unemployment differentials by differences in productivity and tastes. A feminist point of view, on the other hand, suggests that even the portion of the gap explained in this way can be considered a product of another fonn of discrimination, which works:more subtly through the socialization process and affects women's "free" choices about themselves.

Let us consider a hypothetical example of how these two contrasting

approaches would deal with the earnings gap which exists even between men

“and women with apparently similar characteristics. Suppose a male college

graduate is earning one and a half times as much as a female college


Suppose further that these two individuals, aside from being of differentj

sexes, are alike in in all observed characteristics; most importantly, they I’

have both completed the same amount of schooling and have the same number fg',l”“

of years of work experience. How do we explain the substantial difference.

in their earnings? The first approach outlined above would emphasize the@

importance of the remaining productivity differences between these two people,‘

which are less easily measurable. These include quality differences in

schooling and post-schooling training which are not captured in a simple one-dimensional measure such as number of years of schooling or experience. For example, the man might be earning more because he "chose" a major in

This would imply a female/male wage ratio of 0.667.

The actual ratio is somewhat lower than this. In 1970, it was 0.594.

:1’ ._


ml 3 l 3 ': §« 5 x §


.,-bx. mm a... .. M... .«.-..u»ni..en....«


,.:~...«\<’.ar»-u.'»..-s. «no-'\-v.2-1-«>-r~»‘v».:a .a-..v».~ u - -F»~'=

,.‘-4-51 .4». pox

M...‘ s -..\..~  -14.»

F. ~1-—»<-»

mathematics or business administration in college while the woman "chose"

a major in literature. Although they have both worked for the same number of years, perhaps the man's jobs formed an orderly career progression, with substantial investment in on—the—job training, while the woman held a seriesl of dead—end or unrelated jobs and actually received very little in the way of training on the job. The woman's disconnected work history may haven resulted from her anticipation of possible withdrawal from the labor force

or from an attempt to adapt her work lifetto her other family commitments.

This economic approach looks to differences such as these, rather than to{ discrimination against women, in seeking to explain the "unexplained" portion ‘of the earnings gap. Since we can never observe and measure all the relevant variables so precisely, any statement as to what we would find if we couldl do so mu§h necessarily remain unproven.

Other economists have looked at these questions from a feminist perspective and find the point of View described above to be of dubious value, and to ix.

have two important weaknesses. First, even if we could measure and allow ”

for any and all male—female productivity differences stemming from differingnill

education, training and work commitment, it is extremely unlikely that we could ever completely account for the entire gap between the earnings of men and women. Second, these productivity differences themselves are the result of a less direct form of discrimination, frequently referreg to as,role, differentiation, to which both men and women are subject from birth. Women are conditioned to expect to spend a substantial proportion of their lives outside the labor force, and are implicitly or explicitly discouraged from investing heavily in market—oriented skills and career preparation. Actuali di§Rgimination against women in the labor market reinforces this effect,

and helps to make this socially conditioned expectation of secondary status

a Self-fulfilling prophecy,


.,' i r 9 3

\,__A,- ,,__»_-,,i. ‘,,._._.»..y.=....‘ .<~,.-;.-.-.-.,.-~.~L- -‘»—.-..-w vfi,~:ri   '

._ .....¢... . .._. -..~.~.\-1 :.;,~*'



.;. ~w:t'~:.. «raw —.«r- «I -

« Aa-~‘.a.»...  .«~.. ..\: I-....-..~.u...-,.-....;........ ‘Iva: -,n. . <:

' unemployment. Familiarity with the data collection system and the set of if.

'interviews with a randomly selected sample of approximately 47,000 households.'“

—as are persons on layoff waiting for recall and those waiting to start a

-4- Most of the attention given to economic inequality between sexes; 1ik¢ this discussion so far, has concentrated on the earnings gap. A complete understanding of the economic disadvantages to which women are subject

requires that equally careful attention be given to the sex differential

.in unemployment rates. Two aspects of this differential required

explanation: (1) the fact that the female unemployment rate has been consistently higher than the male rate, and (2) the fact that recently this gap has been widening, at the same time that the female labor force

participation rate has been rising.


Obviously, any useful analysis of these phenomena must begin with a

clear understanding of the concepts of labor force participation and _‘ fl"{d


definitions used in measuring these key labor market phenomena is essential.

H‘ ‘J

Data on total population, employment and unemployment are collected

monthly by the U.S. Bureau of the Census, and the results are:publishedA_


in the periodical Employment and Earningsgby the Bureau of Labor Statisticsf

of the U.S. Department of Labor. This information is gathered via personal

Each respondent is questioned about his or her employment status during the previous week. All those who did any work for pay or profit or did as much as 15 hours UF unpaid work in a family enterprise are classified as employed, as are persons who had jobs from which they were temporarily absent because of illness, vacation or other personal reasons. All those who made efforts to find work during the preceding four weeks, but did not have a job and

were available for work during the survey week, are classified as unemployed,"

new job within the next thirty days; The total number of employed and

unemployed workers in the population as a whole is estimated on the basis

‘-'3-sic.-.A.«...,J‘ - V-.-.... ....y : .s'm.4'...‘~.._..&.-..- ..g .:w._...»-..'_'» ....‘ L.-use-3-...'.-. ~x.'_ Au. .«.su-~x q. ‘.$.'.-.~.-...-v-‘\ n-.;.....m-ww-. .w«~:‘-M“ ‘ ‘aim. ‘-»- -»


iof the proportion of all individuals of working age (16 and over) in the‘.

sample found to be employed or unemployed. The total labor force is thefl\ simply derived by adding the estimated total number of empkyed workers and unemployed workers. The labor force participation rate is a percentage derived by dividing the labor force by the population of working age. The unemployment rate tells us what percentage of the labor force is unemployed, and is derived by dividingunemployment by the total labor force. Labor force, employment and unemployment statistics are all available for a

wide range of subgroups within the total population, classified by such

characteristics as age, sex, race and marital status. The table below

‘contains the annual average unemployment rates for men and women since 1960.


Year Men Women 1960 5.4 5.9 1961 6.4 7.2 1962 5.2 6.2 1963 5.2 6.5 1964 4.0 6.2 1965 ' 4.0 5.5 1966 3.2 4.8 1967 3.1 5.2 1968‘ 2.9 4.8 1969 2.8 4.7 1970 4.4 5.9 1971 5.3 6.9 1972 4.9 6.6

Source: Manpower Report of the President (Washington, D.C.: G.P.0., March 1973), Table A-1, p. 127.

The relatively high rate of female unemployment can be at least partly explained in terms of differences in mobility between the sexes. Two types of mobility must be distinguished: (1) vintra-labor force mobility, which is movement from one job to another within the labor force: and (2) inter—labor force mobility, which is movement in and out of the labor force. Too little

intra-labor force mobility and too much inter labor force mob'1't b th " 1 1 y o





9'5 -? \ 3 . 4 1

. '1 I

 — 2


.-.~ ~....“». mu-..' u»....:. -o.r..‘c‘.. .... 


£12.- ;.'..f.. -‘-.;a-.'-74 «V-:w..» 

‘female unemployment rate.

1 I ,, 1 u 6 éa. ~-...o-« 4“

-5- contribute to the high rate of female unemployment. when they become

unemployed, women are less likely than men to be able to move to where their best job opportunities are aglable because of their other family commitmentso Thgggeographic and occupational immobility makes it harder for them to find jobs, and accounts for some of the female~male unemployment differential.

A more important factor is the extensive movement of women into and "V“ a

out of the labor force. Women clearly exhibit more inter-labor force

mobility? and consequently more total mobility than men. Thus at any point


in time a relatively large number of women will have just entered the labor _ 5,gf3

force and be searching for jobs, and the effect of this is to increase the

Direct discrimination against women may also be a significant factor


in determining their unemployment situation. The "over-crowding" hypothesis points out that women have been and continue to be largely segregated in

certain clerical and service occupations. The occupational segregation



of men and women is certainly substantial. For example, in 1960 47% of

‘ V I D U ’ . ‘ ,_,, ,;.,..,,,,,_~ ..,_......; ,4.-..».«.-5/:~

all women workers were employed in occupations in which at least 80% of the.

workers in the occupation were female; while only 2% of employed men were

xi « n-\'.‘9" "“fi—"

in these same occupations. On the other hand, only 20% of employed women were in occupations in which they represented less than 33% of total

employment, while almost 90% of employed men were in these same occupations3V

_ _ . . _ . _ . . . . ....'....-..__.. 5

2. One index of labor force turnover is the ratio of the proportion of the )~‘

population who were in the labor force at any time during the year to

the average labor force participation rate of the populigi ntfgr that

year. Using this measure, labor force turnover for w§menK§élfieen 25-30%

during the lasiififteen years was opposed to roughly, 2% for men.

See Beth Niemi, "The Female—Male Differential In Unemployment Rates",

Industrial and Labor Relations Review, Vol. 27, No. 3 (April l974), pp. 334-37.

.,,. ..»\;.‘- ,.v»., v\‘¢4 « 4z.....<...-x -.....'_~.-

3. Harriet Zellner, "The Determinants of Occupational Segregation," in Cynthia B. Lloyd (ed) Sex, Discrimination and the Division of Labor- (New York: Columbia Univgrs

:~.-PX‘ .«.-.. 4.. Zr. «s..~:.'- »:~...a.—»««~ «

ity Press, forthcoming in 1975).

‘ $\\‘ o—,a}.‘_~ .'-

- unlikely that geographic or occupational immobility has increased over time;'

women's long—term commitment to labor market activity has increased along with

L., 0. -, -.y


The fact that women seeking employment are largely limited to "feminine" occupations narrows their options and raises the rate of female unemployment.

Thus, there are-inn: three factors that raise the unemployment rate

of women: (1) geographic and occupational immobility, (2) excessive inter~labor force movement, and (3) discrimination and occupation segregation. In

weighing the relative importance of these three forces, we should keep in

mind that we should be able to explain not only the high rate of unemployment ’ ' *g‘gjr among women, but also the widening gap between the unemployment rates of men and women as female labor force participation has risen over time.‘ The p S fill labor force participation rate of women, which was only 31.8% in 1947, had ‘ ilgV; ‘E risen to 43.9% in l972,and continues its upward trend? Simultaneously, the

relative unemployment situation of women has worsened. It is highly

.;»;.3 ,.—«.tufl v 3'\ < 4‘ “‘

Beth Niemi, "Geographic Immobility and Labor Force Mobility: A Study of Female Unemployment,” in Cynthia B. Lloyd (ed.) Sex, Discrimination and

the Division of Labor (New York: Columbia University Press, forthcoming in  o

1': ":<‘~.~‘-».;-‘S.-.‘.'.~‘.'.~.‘, .15-‘;‘.\‘;'..aJ.-....f;.'.u..5e..... ~

_, job is necessarily greater. "Statistical discrimination" occurs when an


-.._..«.. -—-A < -

-8- Neither labor force turnover, the changing age distribution of the \ . , 1 . . , . female labor force, nor changes in specific institutions such as unemployment insurance can explain the widening differential between the unemployment rates of men and women, but the trends observed do appear to be consistent the -?orc.& ”over—crowding hypothesis” described above. As female laborkparticipation rises, the occupational segregation of women both depresses their relative

. I _-'I~.-‘~.‘ wages and raises their unemployment rate.~ -' _ I ‘lV~’

A related aspect of the discrimination process which is of particular


concern today is employers‘ expressed reluctance to hire and train women or to ‘d”F“1 pay them equally because of the supposed greater risk of their quitting, »;;j(fY?

Although it is true that aggregate turnover for women is greater than for men “:_i :7.

w_.«. .. ‘

‘this does not imply that the likelihood of a woman's quitting a specific K

'-.;... Jqo..«':1--

employer hires or promotes workers on the basis of inaccurate generalizations A %%L.‘ about the comparative job stability of male and female workers, Actually, women workers have favorable records of attendance and labor turnover when compared with men employed at similar job levels and under similar circumstancee;gfLl:"

Such factors as the skill level of the job, the age of the worker, and the

,,,~..—.~... ...

worker's length of service with the employer are all much more useful in

. .. .-.»_»

understanding differences in job stability than the fact that the worker is'a

. V man or woman.6 '

U.S. Department of Labor, Wage and Labor Standards Administration, , Facts about Women's Absenteeism and Labor Turnover, August 1969, p. 1.

-.x.§+. ..~... 4; .3‘ . 4% -...._. u...../u. ....a....;u.u-e a...“ .»«~.....~.

, , ,, . V . .—.«. ..~.,\.,.'.\.,._.x..c...,..w.;».-.9:-..,~4a««..

' .’ “"3, ‘villi!-.):‘V ,- -4.-J P'<">#-\'."‘I‘\I‘ .r.~,..._._ S’ ‘_. .V _. iv, .‘.‘.lv-_.y 2:‘ V~ _ [Q ‘ __ ‘. >. ‘ /‘V . h‘

»- ..~-..-. _


Another important question has to do with whether discrimination will diminish as women become increasingly attached to the labor force. Many observers are confident that more women on the job with more skills and more continuous work histories will be sufficient to change employers‘ and fellow workers‘ attitudes. Others are more pessimistic. Their pessimism appears to stem from a basic belief that discrimination cannot be entirely h

understood in such objective terms but is more fundamentally engrained in :~7C


Various attitudes about male and female roles.

W  i5

.2... in. «.3.

? ‘i 1 Q s -"5.